LENOVO ThinkPad E420 - frozen with newer kernel. <SOLVED>

[dcung]

I refreshed my mklive ISOs last week - 22/3/2024 (Bullseye & Bookworm).
As usual, I tested the new builds, on few laptops randomly.

One laptop encountered issue (the rest were okay).
It freezes during booting. If it completed the boot, it would freeze shortly after.
It's around the time when it displays 'keyboard shortcuts' or conky. Then I have to power off with holding power button. This is both for Bullseye & Bookworm.

2024-03-27-090724_848x467_scrot.png (276.77 KiB) Viewed 390 times

.
.
.
On this multiboot laptop, I have (Note - These are working, no problem)
- Windows 10
- Bullseye - k-5.10.0-25-amd64.squashfs (Live system built on x86_64 for amd64 on Fri Sep 8 03:50:44 AEST 2023)
- Bookworm - k-6.1.0-13-amd64.squashfs (Live system built on x86_64 for amd64 on Mon Oct 9 04:57:03 AEDT 2023)

FYI - just to satisfy myself, I installed another blank/formatted SSD, so I can start fresh.
I found that any Bullseye/Bookworm build later that the one mentioned above have the same issue.
I have done few mklive refreshed since then, so I tested on this laptop - XFCE4, JWM, DDOG - all frozen, same symptoms.
For Bullseye : k-5.10.0-26, k-5.10.0-27, k-5.10.0-28
For Bookworm : k-6.1.0-16, k-6.1.0-17, k-6.1.0-18
EDIT: k-6.1.0-16 seem to be working.

I also found that, I can start fresh with
Bullseye - k-5.10.0-25-amd64.squashfs (Live system built on x86_64 for amd64 on Fri Sep 8 03:50:44 AEST 2023)
Bookworm - k-6.1.0-13-amd64.squashfs (Live system built on x86_64 for amd64 on Mon Oct 9 04:57:03 AEDT 2023)

do

Code: Select all

apt update; apt upgrade

Then change to any later kernel files and it works.
So, I will make remaster of these for the time being as my workaround, until I figure out why.

HW info - Extract from dmidecode, etc.
LENOVO ThinkPad E420 Model 1141CTO
Intel(R) Core(TM) i5-2520M CPU @ 2.50GHz
250GB SSD, 4GB RAM, GPU - Intel HD Graphics 3000
BIOS Information
Vendor: LENOVO
Version: 8HET44WW(1.26)
Release Date: 08/07/2013

Out of curiosity, I also tested EasyOS v5.6.7 - Frozen.

[dcung]

Since I became aware of this issue last week, I have been googling and reading other peoples' experiences with E420 (and/or Thinkpad on general).
I have found no concrete 'solution'/answer.

Of interest to me is thinkfan. Some people install it to control noise of their fan, some install it cause they experience over heating.
Bottom line - they found the default setting (in BIOS?) is unsuitable for them. On mine, I found no evidence of overheating and it's reasonably silent.
(BTW - not long ago I have dissembled, clean fan, put new heat paste, heat pad on this laptop)

https://github.com/vmatare/thinkfan
https://forum.thinkpads.com/viewtopic.php?t=119546

I found subtle differences when I use different kernel files in changes dir-content it created (fresh install), in regard to acpi files, etc.
I didn't take notes or pay close attention to be certain though.
But I'm guessing this is why I can use 'changes' from Bullseye k-5.10.0-25-amd64 or Bookworm k-6.1.0-13-amd64 and replace with newer kernel and it still works.
I am guessing that actual perfect E420 config does not exist yet, but at least in k-5.10.0-25-amd64/k-6.1.0-13-amd64 created config it didn't cause frozen like newer ones. These 'missing'/unsuitable config that govern when the fan rev up/down caused heat spikes and freeze the laptop. Just my own hypothesis .

Output of sensors command indicated that at least the CPU fan reading is incorrect (disabled?). Always show 0 RPM.
My fan (and sensors) is working, as I can hear it revving up (and feel more warm air) when I put load on the laptop.

Without load

Code: Select all

root@live:~# sensors
coretemp-isa-0000
Adapter: ISA adapter
Package id 0:  +56.0°C  (high = +86.0°C, crit = +100.0°C)
Core 0:        +53.0°C  (high = +86.0°C, crit = +100.0°C)
Core 1:        +51.0°C  (high = +86.0°C, crit = +100.0°C)

BAT0-acpi-0
Adapter: ACPI interface
in0:          12.35 V  

acpitz-acpi-0
Adapter: ACPI interface
temp1:        +39.0°C  (crit = +120.0°C)

thinkpad-isa-0000
Adapter: ISA adapter
fan1:           0 RPM

root@live:~# 

With full load - I ran mklive build to force 100% CPU.

Code: Select all

root@live:~# sensors
coretemp-isa-0000
Adapter: ISA adapter
Package id 0:  +85.0°C  (high = +86.0°C, crit = +100.0°C)
Core 0:        +85.0°C  (high = +86.0°C, crit = +100.0°C)
Core 1:        +83.0°C  (high = +86.0°C, crit = +100.0°C)

BAT0-acpi-0
Adapter: ACPI interface
in0:          12.35 V  

acpitz-acpi-0
Adapter: ACPI interface
temp1:        +39.0°C  (crit = +120.0°C)

thinkpad-isa-0000
Adapter: ISA adapter
fan1:           0 RPM

root@live:~# 

In the attempt to 'correct' the sensors output, I did attempt to install/compile/tinker with thinkfan - not successful.
Probably beyond my skillset

Anyway, I'm giving up investigating it further.
I'm just gonna use the 'workaround' if I need later kernel.
I know what works and what doesn't on this laptop.
I have tested different frugal install since it took little effort to just drag and drop, and reboot.

For now Windows 10, Bullseye, Bookworm should be sufficient.

[dcung]

I just couldn't let it go.

I did more testing. The general indication is that older Dog mklive Build work.
EasyOS v541, v567, v570 - frozen.

I did marathon mkbuild (using Bookworm with k-6.1.0-16-amd64 - last kernel version that boot fresh w/o problem) to see if heat is a problem. The temp gets high, the fan revved up when CPU is 100% when filesystem.squashfs is created. But the laptop never froze - script making 20 builds non-stop.
So, I figured, heat is not the problem. Although, later I have to investigate later why sensors command always show 0 RPM.

HOT - 80 to 92 when nearing the 20th builds.

Code: Select all

root@lenovo-e420:~# sensors
coretemp-isa-0000
Adapter: ISA adapter
Package id 0:  +92.0°C  (high = +86.0°C, crit = +100.0°C)
Core 0:        +92.0°C  (high = +86.0°C, crit = +100.0°C)
Core 1:        +90.0°C  (high = +86.0°C, crit = +100.0°C)

BAT0-acpi-0
Adapter: ACPI interface
in0:          12.43 V  

acpitz-acpi-0
Adapter: ACPI interface
temp1:        +39.0°C  (crit = +120.0°C)

thinkpad-isa-0000
Adapter: ISA adapter
fan1:           0 RPM
fan2:           0 RPM

root@lenovo-e420:~# 

Google searching, I read that some people have issues with E420 caused by i915 driver. And I remember @fredx181 had issue with it before. I searched and found the discussion.
viewtopic.php?p=90188#p90188

I tried it, and it works!
I only tried DDOG of my latest Bullseye & Bookworm refresh that would not boot on this laptop before.

Code: Select all

title E420/bullseye-amd64-DDog.k28  -  Save changes
 uuid e9f7b374-8ad9-4cc8-94e8-d77102e4a08b
 kernel /E420/bullseye-amd64-DDog.k28/live/vmlinuz1 from=/E420/bullseye-amd64-DDog.k28/ noauto changes=/E420/bullseye-amd64-DDog.k28/ i915.enable_dc=0
 initrd /E420/bullseye-amd64-DDog.k28/live/initrd1.xz

title E420/bookworm-amd64-DDog.k18  -  Save changes
 uuid e9f7b374-8ad9-4cc8-94e8-d77102e4a08b
 kernel /E420/bookworm-amd64-DDog.k18/live/vmlinuz1 from=/E420/bookworm-amd64-DDog.k18/ noauto changes=/E420/bookworm-amd64-DDog.k18/  i915.enable_dc=0
 initrd /E420/bookworm-amd64-DDog.k18/live/initrd1.xz

[dcung]

Well.
Spoke too early.

Didn't seem to make much difference when I test longer, still frozen intermittently.

I guess, this is in the category of old HW, too hard to support...
Just use what works.

Anyway, this is a curiosity, not a necessity project

[dcung]

I figured out what is happening - thanks to CPU-X.
It turns out this E420 has dual GPU - Intel and Radeon.
Once I disabled one, and not "Allow OS to switch GPU" in BIOS - problem goes away.

2024-04-15-gpu.png (122.84 KiB) Viewed 78 times